Scaling an engineering organisation isn't primarily a technical problem. It's a systems problem — and the system includes people, processes, and culture just as much as it includes code.
After 15 years of building and leading engineering teams across fintech, iGaming, and logistics, here's what I keep coming back to.
Hire for the problem, not the stack
The best engineers I've worked with weren't defined by the languages they knew. They were defined by how they thought about problems. When you're scaling fast, you need people who can reason about trade-offs, not people who happen to know your ORM.
This doesn't mean technical skills don't matter — they do. But the hierarchy is: problem-solving first, system thinking second, specific tech third.
Process should reduce friction, not create it
Every process you introduce has a cost. The question isn't "is this best practice?" — it's "does this make us faster or slower right now?"
I've seen teams drown in ceremony that made sense at a previous scale. Stand-ups that run 45 minutes. Sprint planning that takes half a day. Retrospectives that produce action items nobody tracks.
Good process at scale looks like:
- Short feedback loops — deploy often, measure everything, fix fast
- Clear ownership — every service, every domain has a name attached
- Minimal coordination overhead — if two teams need a meeting to ship a feature, your architecture is wrong
Architecture follows team structure
Conway's Law isn't just an observation — it's a design tool. If you want independent, fast-moving teams, your architecture needs to support that. Shared databases, shared deployment pipelines, shared anything becomes a bottleneck.
The most effective replatforming work I've done started with the org chart, not the system diagram.
Observability is not optional
You cannot scale what you cannot see. Before optimising anything — performance, cost, reliability — you need to know what's actually happening in production.
This means:
- Structured logging from day one
- Distributed tracing across service boundaries
- Dashboards that answer questions, not just display metrics
- Alerts that are actionable, not noisy
I've walked into organisations running 100+ services with no centralised logging. The first conversation is always the same: "We need to see before we can fix."
The work is never purely technical
The hardest part of scaling engineering isn't the code. It's getting alignment between engineering, product, and the business on what matters. It's making trade-offs visible. It's saying "no" to the right things so you can say "yes" to the things that actually move the needle.
That's the work I find most interesting — and it's why I do what I do.
If you're dealing with scaling challenges and want to talk through your situation, get in touch.